Structural crack is an important factor which causes failure of reinforced concrete bridges. In this work, automatic detection and dimensional measurement of concrete bridge crack are researched, for improving technical level and efficiency of concrete bridge state assessment. Images containing crack features are first recognized using information entropy characteristics of intensity clustering, for promoting efficiency and robustness of rough crack localization based on proportional segmentation. After the features are refined at sub-pixel level, their actual dimensions are accurately measured employing a cross structured light system. Experiments show that the problems such as high misjudgment, low efficiency and poor accuracy in the existing technologies are preliminarily addressed; the proposed method performs well in crack detection and measurement using concrete bridge structure images.
This paper presents an automatical crack recognition approach. Compared with the existing methods, it has a significant increase in robustness and efficiency when faced with widely varying field conditions. Inherent characteristics of crack images are exploited using proportional segmentation, combined with robust feature extraction to improve machine learning classifier performance. Experiments show that this method perform well in crack images recognition across different concrete conditions.
A key technology of multi-camera visual measurement system is global calibration. The global calibration methods existed either has disadvantages such as high cost and complicated operation, or has limited application areas only for visual system based on stereo vision. A convenient global calibration method for multi-camera system based on two planar targets is proposed in this paper, and the pose relation between the two targets can be unknown. This method is not only suitable for system based on stereo vision, but also for system based on monocular vision. It has been used in the factory-calibration of four-wheel aligner consisted of 2 or 4 cameras, and calibration precision meet the requirement.
In order to solve the robustness and efficiency problems of chessboard corner detection under on-site condition, a method based on the square-closed loop template and local gray symmetry factor is proposed for detecting sub-pixel chessboard corners with high precision and efficiency. Interest points on edges of original image is detected by square-closed loop template according to the transition times on the template, on this basis, corners are roughly detected by averaging the adjacent coordinates of interest points; gray symmetry factors, calculated in the local neighborhood of roughly detected corners, are used as the weighting factors to precisely detect corners on sub-pixel level. Experimental results indicate that computing speed and positioning accuracy of this method was obviously increased. The corner detection performance could be significantly improved using the proposed method.
A model was built for four-wheel alignment on a vehicle based on computer vision. Such parameters as toe-in angle, camber angle, kingpin inclination and kingpin caster were accurately defined and calculation formulas were formulated for the parameters, especially the calculation methods for vector N and vector E. A kind of new 3D four-wheel aligner vas developed. Simulation results and actual measurements indicated that the model and solution method were feasible and effective.
Positioning parameters of four-wheel have significant effects on maneuverabilities, securities and energy saving abilities of automobiles. Aiming at this issue, the error factors of 3D four-wheel aligner, which exist in extracting image feature points, calibrating internal and exeternal parameters of cameras, calculating positional parameters and measuring target pose, are analyzed respectively based on the elaborations of structure and measurement principle of 3D four-wheel aligner, as well as toe-in and camber of four-wheel, kingpin inclination and caster, and other major positional parameters. After that, some technical solutions are proposed for reducing the above error factors, and on this basis, a new type of aligner is developed and marketed, it’s highly estimated among customers because the technical indicators meet requirements well.
Surface of a precision rotor, a typical helical surface with complicated 3-D shape, is hard to measure with CMM accurately due to the difficulties of spiral scan and probe compensation. Aiming at this problem, an apparatus, avoiding probe compensation directly by means of non-contact measuring, is designed and developed based on laser triangulation, and the corresponding method for measuring parameter lead and transverse section profile is proposed based on geometric feature model expressing the surface of precision rotor. After apparatus is calibrated with a standard plane and ring gauge, experiments are carried out to scan spiral line and measure transverse section profile respectively, and results of lead error detection and profile error evaluation satisfactorily match the theoretical values provided by manufacturer.
At present, probe compensation is the key problem in measuring geometric parameters of complex screw surface with CMM due to its complicated 3D shape, aiming at this problem, some new measurement methods are proposed based on geometric feature models, expressing the screw surface and its offset surface separately. Supposing the parameter lead of a screw surface is known, it’s realized by scanning one single profile to complete probe compensation and calculate out all parameters, and the probe compensation is done by two improved methods, named as modified cross product and offset surface virtual measurement respectively, the theory and detailed process of which are discussed in this paper. After performing systematic experiments of profile scan, probe compensation and error evaluation, results show that the new measurement methods provide higher precision, stability and realizability.